U.S. patent application number 15/859135 was filed with the patent office on 2019-03-21 for isolated dc-dc converter.
The applicant listed for this patent is Texas Instruments Incorporated. Invention is credited to Giacomo Calabrese, Giovanni Frattini, Maurizio Granato.
Application Number | 20190089262 15/859135 |
Document ID | / |
Family ID | 65720775 |
Filed Date | 2019-03-21 |
![](/patent/app/20190089262/US20190089262A1-20190321-D00000.png)
![](/patent/app/20190089262/US20190089262A1-20190321-D00001.png)
![](/patent/app/20190089262/US20190089262A1-20190321-D00002.png)
![](/patent/app/20190089262/US20190089262A1-20190321-D00003.png)
![](/patent/app/20190089262/US20190089262A1-20190321-D00004.png)
![](/patent/app/20190089262/US20190089262A1-20190321-D00005.png)
![](/patent/app/20190089262/US20190089262A1-20190321-D00006.png)
![](/patent/app/20190089262/US20190089262A1-20190321-D00007.png)
![](/patent/app/20190089262/US20190089262A1-20190321-D00008.png)
![](/patent/app/20190089262/US20190089262A1-20190321-D00009.png)
![](/patent/app/20190089262/US20190089262A1-20190321-D00010.png)
United States Patent
Application |
20190089262 |
Kind Code |
A1 |
Granato; Maurizio ; et
al. |
March 21, 2019 |
ISOLATED DC-DC CONVERTER
Abstract
An isolated DC-DC converter 400 includes a plurality of modules,
an output node, and a transformer with a secondary side having
first and second terminals and a secondary side ground. Ones of the
modules includes a first low-side switch coupled between the first
terminal and a low-voltage node, and a second low-side switch is
coupled between the second terminal and the low-voltage node. A
first high-side switch is coupled between the first terminal and a
high-voltage node, and is activated by a voltage at the second
terminal. A second high-side switch is coupled between the second
terminal and the high-voltage node, and is activated by a voltage
at the first terminal. The isolated DC-DC converter also includes a
switch controller to open and close the first and second low-side
switches at times corresponding to a frequency of the
converter.
Inventors: |
Granato; Maurizio; (Milano,
IT) ; Calabrese; Giacomo; (Freising, DE) ;
Frattini; Giovanni; (Travaco? Siccomario (PV), IT) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
Texas Instruments Incorporated |
Dallas |
TX |
US |
|
|
Family ID: |
65720775 |
Appl. No.: |
15/859135 |
Filed: |
December 29, 2017 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
62560647 |
Sep 19, 2017 |
|
|
|
Current U.S.
Class: |
1/1 |
Current CPC
Class: |
H02M 3/33592 20130101;
H02M 1/088 20130101; H02M 2001/0058 20130101; H02M 2001/0054
20130101; H02M 3/337 20130101 |
International
Class: |
H02M 3/335 20060101
H02M003/335; H02M 1/088 20060101 H02M001/088 |
Claims
1. An isolated DC-DC converter, comprising: an output node for
delivering an output voltage; a transformer including a secondary
side having first and second terminals and a secondary side ground;
a plurality of modules, respective ones of the modules including:
first and second low-side switches, the first low-side switch
coupled between the first terminal and a low-voltage node, and the
second low-side switch coupled between the second terminal and the
low-voltage node, wherein a first voltage is across the first
low-side switch, and a second voltage is across the second low-side
switch; and first and second high-side switches, the first
high-side switch coupled between the first terminal and a
high-voltage node and configured to be activated by a voltage at
the second terminal, and the second high-side switch coupled
between the second terminal and the high-voltage node and
configured to be activated by a voltage at the first terminal; and
a switch controller configured to open and close the first and
second low-side switches at times corresponding to a frequency of
the converter, by: closing the first low-side switch when the first
voltage is zero; during a first phase, opening the first low-side
switch when the first voltage is zero and a current through the
secondary side is zero; closing the second low-side switch when the
second voltage is zero; and during a second phase, opening the
second low-side switch when the second voltage is zero and the
current through the secondary side is zero; a first one of the
modules being an output module whose high-voltage node includes the
output node; a second one of the modules being a ground module
whose low-voltage node includes the secondary side ground; and the
high-voltage node of the ground module being coupled to the
low-voltage node of the output module.
2. The converter of claim 1, further comprising multiple isolation
capacitors connected to isolate connections between the ground
module and the secondary side from connections between the output
module and the secondary side.
3. The converter of claim 1, wherein the output node is configured
to have twice the voltage of the low-voltage node of the output
module and of the high-voltage node of the ground module.
4. The converter of claim 1, wherein an input node of the converter
is configured to receive an input voltage less than or equal to
five volts above a voltage of a ground of a primary side of the
transformer, and wherein the output node is configured to deliver
the output voltage greater than five volts above a voltage of the
secondary side ground and double the input voltage.
5. The converter of claim 4, further comprising: a first output
capacitor coupled to the output node, and coupled to the output
module low-voltage node and the ground module high-voltage node;
and a second output capacitor coupled to the output module
low-voltage node and the ground module high-voltage node, and
coupled to the secondary side ground.
6. The converter of claim 5, wherein the first and second output
capacitors are each rated for use at five volts or less.
7. The converter of claim 1, wherein the low-side switches include
NMOS transistors and the high-side switches include PMOS
transistors.
8. The converter of claim 1, further comprising: an input node for
receiving an input voltage; the transformer further including a
primary side having first and second primary terminals and a
primary side ground; first and second primary low-side switches,
the first primary low-side switch coupled between the first primary
terminal and the primary side ground, and the second primary
low-side switch coupled between the second primary terminal and the
primary side ground, wherein a first primary voltage is across the
first primary low-side switch, and a second primary voltage is
across the second primary low-side switch; first and second primary
high-side switches, the first primary high-side switch coupled
between the first primary terminal and the input node and
configured to be activated by a voltage at the second primary
terminal, and the second primary high-side switch coupled between
the second primary terminal and the input node and configured to be
activated by a voltage at the first primary terminal; and to cause
the first and second primary voltages to alternatingly be zero, by:
closing the first primary low-side switch when the first primary
voltage is zero; during a first primary phase, opening the first
primary low-side switch at a time to enable a current through the
primary side to charge the first primary terminal to the input
voltage while discharging the second primary terminal to reduce the
second primary voltage to zero; closing the second primary low-side
switch when the second primary voltage is zero; and during a second
primary phase, opening the second primary low-side switch at a time
to enable the current through the primary side to charge the second
primary terminal to the input voltage while discharging the first
primary terminal to reduce the first primary voltage to zero.
9. An isolated DC-DC converter, comprising: an output node for
delivering an output voltage; a transformer including a secondary
side having two terminals and a secondary side ground; and a ground
module and an output module, respective ones of the modules
including: a low-side circuitry for driving the secondary side
configured to: connect alternating ones of the terminals to a
low-voltage node at times corresponding to a frequency of the
converter such that a voltage at a selected terminal to be
connected is a voltage of the low-voltage node, and disconnect the
selected terminal when a current through the secondary side is zero
and a voltage between the selected terminal and the low-voltage
node is zero; and a high-side circuitry for driving the secondary
side configured to: when a node coupled to the selected terminal is
charged to a voltage of a high-voltage node, disconnect a
non-selected terminal from the high-voltage node, and when a node
coupled to the non-selected terminal is discharged to the
low-voltage node voltage, connect the selected terminal to the
high-voltage node; wherein the high-voltage node of the output
module includes the output node, the low-voltage node of the ground
module includes the secondary side ground, and the high-voltage
node of the ground module is coupled to the low-voltage node of the
output module.
10. The converter of claim 9, further comprising multiple isolation
capacitors connected to isolate connections between the ground
module and the secondary side from connections between the output
module and the secondary side.
11. The converter of claim 9, wherein the output node is configured
to have twice the voltage of the low-voltage node of the output
module and of the high-voltage node of the ground module.
12. The converter of claim 9, wherein: an input node of the
converter is configured receive an input voltage less than or equal
to five volts above a voltage of a ground of a primary side of the
transformer, and the output node is configured to deliver the
output voltage greater than five volts above a voltage of the
secondary side ground and double the input voltage.
13. The converter of claim 12, further comprising: a first output
capacitor coupled to the voltage output, and coupled to the output
module low-voltage node and the ground module high-voltage; and a
second output capacitor coupled to the output module low-voltage
and the ground module high-voltage node, and coupled to the
secondary side ground.
14. The converter of claim 13, wherein the first and second output
capacitors are each rated for use at five volts or less.
15. The converter of claim 9, wherein the low-side circuitry
includes NMOS transistors and the high-side circuitry includes PMOS
transistors.
16. The converter of claim 9, further comprising: an input node for
receiving an input voltage; the transformer further including a
primary side having two primary terminals and a primary side
ground; a primary low-side circuitry for driving the primary side
configured to: connect alternating ones of the primary terminals to
the primary side ground when a voltage between a selected primary
terminal to be connected and the primary side ground is zero, and
disconnect the selected primary terminal from the primary side
ground at a time to enable the primary current to then charge the
selected primary terminal to the input voltage, and discharge a
non-selected primary terminal to a voltage of the primary side
ground; and a primary high-side circuitry for driving the primary
side configured to: when the selected primary terminal is charged
to the input voltage, disconnect the non-selected primary terminal
from the input node, and when the other primary terminal is
discharged to the primary side ground voltage, connect the selected
primary terminal to the input node.
17. An isolated DC-DC converter, comprising: a transformer
including a primary side having first and second primary terminals
and a primary side ground, and a secondary side having two
secondary terminals and a secondary side ground; and a primary side
module, a secondary side output module and a secondary side ground
module, respective ones of the modules including: two PMOS
transistors with sources connected to a high-voltage node, the PMOS
transistors cross-coupled to corresponding ones of the primary or
secondary terminals; and first and second NMOS transistors with
sources connected to a low-voltage node and with drains of
different ones of the NMOS transistors connected to different
corresponding ones of the primary or secondary terminals; wherein
the corresponding terminals, the high-voltage node, and the
low-voltage node of the primary side module are the primary
terminals, an input node connected to receive an input voltage, and
the primary side ground, respectively; wherein the corresponding
terminals, the high-voltage node and the low-voltage node of the
secondary side output module are the secondary terminals, an output
node connected to deliver an output voltage and the high-voltage
node of the secondary side ground module, respectively; wherein the
corresponding terminals, the high-voltage node and the low-voltage
node of the secondary side ground module are the secondary
terminals, the low-voltage node of the secondary side output module
and the secondary side ground, respectively; wherein the drain of
the first NMOS transistor is coupled to the first primary terminal,
and the drain of the second NMOS transistor is coupled to the
second primary terminal, wherein a first voltage is across the
first primary NMOS transistor, and a second voltage is across the
second primary NMOS transistor; and wherein the first and second
NMOS transistors of the primary side module are configured to
activate and deactivate so that the first and second voltages are
alternatingly zero, by: activating the first primary NMOS
transistor when the first voltage is zero; during a first phase,
deactivating the first primary NMOS transistor at a time to enable
a current through the primary side to charge the first primary
terminal to the input voltage while discharging the second primary
terminal to reduce the second voltage to zero; activating the
second primary NMOS transistor when the second voltage is zero; and
during a second phase, deactivating the second primary NMOS
transistor at a time to enable the current through the primary side
to charge the second primary terminal to the input voltage while
discharging the first primary terminal to reduce the first voltage
to zero.
18. The converter of claim 17, wherein the NMOS transistors of the
output module are configured to be alternatingly activated at times
corresponding to a frequency of the converter such that a voltage
across a selected output module NMOS transistor to be activated is
zero, and deactivated when a current through the secondary side is
zero and a voltage across the selected output module NMOS
transistor is zero.
19. The converter of claim 17, wherein the NMOS transistors of the
ground module are configured to be alternatingly activated at times
corresponding to a frequency of the converter such that a voltage
across a selected ground module NMOS transistor to be activated is
zero, and deactivated when a current through the secondary side is
zero and a voltage across the selected ground module NMOS
transistor is zero.
20. The converter of claim 17, wherein gates of the NMOS
transistors of the ground module are connected to the secondary
side ground.
Description
CROSS-REFERENCES TO RELATED APPLICATIONS
[0001] This application is a non-provisional of and claims priority
to U.S. Provisional Pat. App. No. 62/560,647, filed on Sep. 19,
2017, which is hereby incorporated by reference.
[0002] This application is related to co-pending and co-assigned
U.S. patent application Ser. No. ______ (docket TI-77168), filed on
even date, and entitled, "ISOLATED DC-DC CONVERTER;" and to
co-pending and co-assigned U.S. patent application Ser. No. ______
(docket TI-78124), filed on even date, and entitled, "ISOLATED
DC-DC CONVERTER."
BACKGROUND
[0003] This relates generally to isolated DC-DC converters, and
more particularly to efficient, high input and/or output voltage
isolated DC-DC converters.
[0004] Isolated DC-DC converters typically include a transformer,
with: (a) a power generating input side, typically called the
primary side; and (b) a power output side, typically called the
secondary side. The differences between an isolated converter and a
non-isolated converter are that: (a) in an isolated converter, the
primary side and the secondary side have different grounds; and (b)
in a non-isolated converter, the primary side and the secondary
side share a same ground (such as using an inductor, instead of a
transformer, to convert voltage).
[0005] Isolated DC-DC converters can be small enough to fit within
an integrated circuit ("IC") package, but may require significant
current on the primary side to drive the secondary side across the
isolation barrier, i.e., the gap between the transformer coils.
Such converters can be used for transferring power from a high
voltage domain to a low voltage domain, such as from a power line
to a set of consumer electronics, or from a battery on an electric
vehicle to electronics on that vehicle. Grounds in high voltage
regimes (e.g., the power line or battery) may have voltage swings
(such as 1000V or 1500V swings), which the devices connecting to
those grounds are designed to tolerate. Grounds in the low voltage
regimes may have much smaller swings (such as much less than 1V),
and devices connected to those low voltage grounds could be
rendered inoperative if subjected to the voltage swings tolerated
in a high voltage regime. Isolation barriers separate high voltage
regimes from low voltage regimes, so that power transfer can be
performed without damaging low voltage components.
SUMMARY
[0006] In described examples, an isolated DC-DC converter includes:
an output node for delivering an output voltage; a transformer
comprising a secondary side having first and second terminals and a
secondary side ground; and a plurality of modules. Ones of the
modules include respective: first and second low-side switches, and
first and second high-side switches. The first low-side switch is
coupled between the first terminal and a low-voltage node, and the
second low-side switch is coupled between the second terminal and
the low-voltage node. A first voltage is across the first low-side
switch, and a second voltage is across the second low-side switch.
The first high-side switch is coupled between the first terminal
and a high-voltage node, and is configured to be activated by a
voltage at the second terminal. The second high-side switch is
coupled between the second terminal and the high-voltage node, and
is configured to be activated by a voltage at the first terminal.
The isolated DC-DC converter also includes a switch controller
configured to open and close the first and second low-side switches
at times corresponding to a frequency of the converter, as follows.
The first low-side switch is closed when the first voltage is zero.
During a first phase, the first low-side switch is opened when the
first voltage is zero and a current through the secondary side is
zero. The second low-side switch is closed when the second voltage
is zero. During a second phase, the second low-side switch is
opened when the second voltage is zero and the current through the
secondary side is zero. A first one of the modules is an output
module whose high-voltage node includes the output node. A second
one of the modules is a ground module whose low-voltage node
includes the secondary side ground. The high-voltage node of the
ground module is coupled to the low-voltage node of the output
module.
BRIEF DESCRIPTION OF THE DRAWINGS
[0007] FIG. 1 shows an example of a transistor-level diagram of a
circuit portion used to construct an isolated DC-DC converter.
[0008] FIG. 2 shows a transistor-level diagram of an isolated DC-DC
converter.
[0009] FIG. 3 shows a timing diagram for an isolated DC-DC
converter as shown in and described with respect to FIG. 2.
[0010] FIG. 4 shows a transistor-level diagram of a high output
voltage isolated DC-DC converter.
[0011] FIG. 5 shows a transistor-level diagram of a high output
voltage isolated DC-DC converter.
[0012] FIG. 6A shows an example of a process for controlling an
isolated DC-DC converter.
[0013] FIG. 6B shows an example of a process for controlling an
isolated DC-DC converter.
[0014] FIG. 7 shows an example of a transistor-level diagram of a
circuit portion used to construct an isolated DC-DC converter.
[0015] FIG. 8 shows a transistor-level diagram of an isolated DC-DC
converter.
[0016] FIG. 9 shows an example of a timing diagram for an isolated
DC-DC converter as shown in FIG. 8.
[0017] FIG. 10 shows an example of a transistor-level diagram of a
high voltage isolated DC-DC converter.
DETAILED DESCRIPTION OF EXAMPLE EMBODIMENTS
[0018] Example embodiments include methods, devices and systems for
efficient (and in some embodiments, high input-output voltage),
isolated DC-DC converters, particularly using circuits and
components that can be incorporated into an integrated circuit (IC)
package. In some embodiments, the converters are capable of
operating at large minimum voltage differentials between Vin (input
voltage) and primary ground, and/or between Vout (output voltage)
and secondary ground (e.g., 10V or 15V), as further described
hereinbelow.
[0019] FIG. 1 shows an example of a transistor-level diagram of a
circuit portion 100 used to construct an isolated DC-DC converter.
As shown in FIG. 1, an input voltage Vin 102 is connected to the
sources of two PMOS transistors, P1 104 and P2 106 (PMOS conduction
type). (Diodes shown are intrinsic diodes of respective PMOS and
NMOS transistors.) Nodes V1 108 and V2 110 are cross-coupled
between the drains and gates of P1 104 and P2 106 (high side
switches). Accordingly, node V1 108 is connected to the drain of P1
104 and is connected to the gate of P2 106; and node V2 110 is
connected to the drain of P2 106 and is connected to the gate of P1
104.
[0020] An inductor 112 can also be considered a primary side or a
secondary side of a transformer--accordingly, the primary side or
secondary side of a transformer is an inductor, and circuit portion
100 can be used to construct an isolated DC-DC converter such that
the primary side or secondary side of a transformer in the
converter stands in place of the inductor 112 (see, e.g., FIGS. 2,
4, 5, 8 and 10). The inductor 112 is connected between nodes V1 108
and V2 110.
[0021] Parasitic capacitor C1 114 and the source/drain path of NMOS
transistor N1 116 (NMOS conduction type) are connected in parallel
between node V1 108 and GND 118 (the source of N1 116 is connected
to GND 118). C1 114 corresponds to (represents) parasitic
capacitances of N1 116, P1 104 and P2 106. For example, C1 114
includes drain-source capacitance of N1 116, drain-source
capacitance of P1 104, and gate-source capacitance of P2 106. The
gate of N1 116 is connected to a (buffered) control signal G1
120.
[0022] Parasitic capacitor C2 122 and the source/drain path of NMOS
transistor N2 124 are connected in parallel between node V2 110 and
GND 118 (the source of N2 124 is connected to GND 118). C2 122
corresponds to (represents) parasitic capacitances of N2 124, P1
104 and P2 106. For example, C2 122 includes drain-source
capacitance of N2 124, drain-source capacitance of P2 106, and
gate-source capacitance of P1 104. The gate of N2 124 is connected
to a (buffered) control signal G2 126. (N1 116 and N2 124 comprise
low side switches.)
[0023] Inputs G1 120 and G2 126 are preferably generated by a
control signal generator 128. An isolated DC-DC converter (e.g., as
in FIGS. 2, 4, 5, 8 and 10) preferably comprises a primary control
signal generator and a secondary control signal generator, with the
generators respectively and separately connected to generate and
deliver control signals to drive controlled switches (e.g., N1 116
and N2 124).
[0024] Preferably, power losses are minimized. Converters can lose
power, such as through switching losses, conduction losses, and
other dissipative losses. Switching losses in MOSFET devices come
from the dynamic voltages and currents a MOSFET handles during the
finite, nonzero time it takes to turn on or off. Conduction losses
are power dissipation losses resulting from current moving through
a component; for example, MOSFET devices have nonzero resistance.
Other dissipative losses occur, such as when power is discharged to
ground, e.g., when a charged gate or capacitor is connected to
ground, resulting in the gate or capacitor discharging.
[0025] In operation, the NMOS transistors N1 116 and N2 124 are
preferably zero voltage switched (ZVS) and zero current switched
(ZCS). Accordingly, G1 120 and G2 126 are respectively controlled
to turn N1 116 and N2 124 on OR off (generally one or the other and
not both; the other of turn-on or turn-off timing is determined
using other criteria, such as further described with respect to
FIG. 3), when corresponding adjacent nodes V1 108 and V2 110 are
fully discharged (at zero voltage with respect to GND 118), and
zero current flows across the inductor 112 (from node V1 108 to
node V2 110 or vice versa). This results in minimal switching
losses for either the turn-on or turn-off transition, because
little or no voltage is across (and little or no current is
conducted by) N1 116 and N2 124 while they are switched. Zero
voltage switching and zero current switching are further described
hereinbelow with respect to FIG. 3. (Where this description refers
to zero voltage at a node, rather than across a component, it means
that the node is at a voltage of a corresponding ground of a
primary side or secondary side of a transformer.)
[0026] In embodiments as shown in FIG. 1, power cost for MOSFET
control is reduced by making P1 104 and P2 106 self-driving.
Accordingly, rather than requiring control signals to bias P1 104
and P2 106, P1 104 and P2 106 are biased by charge driven to P1 104
and P2 106 via nodes V2 110 and V1 108 (respectively) by the
current through the resonance (inductor 112). Self-driving P1 104
and P2 106 reduces the complexity and power cost of controlling the
MOSFETS.
[0027] Further power savings are achieved by using the current
across the inductor 112 (the resonance current) to drive charge out
of the gates of MOSFETS P1 104 and P2 106 and turn that charge into
current across the inductor 112. Some of that current charges
capacitances C2 122 and C1 114. (Capacitances C1 114 and C2 122
include gate-source parasitic capacitances of P1 106 and P2 104,
respectively.) This recycling avoids dissipation of gate charges
(Cgs charges) when P1 104 and P2 106 are switched, because those
gate charges are reused. Accordingly, those gate charges (charges
of gate-source parasitic capacitances), along with charges of
respective drain-source and other parasitic capacitances, are
driven out of C1 114 or C2 122 by the resonance current, turned
into current across the inductor 112, and (in part) are used to
charge C2 122 or C1 114 (respectively).
[0028] Because gate charge of P1 104 and P2 106 are reused, the
tradeoff between QG (gate charge) and RDS (drain-source on
resistance) can be largely avoided for P1 104 and P2 106. For
example, a lower RDS can generally be used, thereby enabling faster
switching and lower heat output, without resulting in excessive
power costs due to a higher QG, because the gate charge power cost
is largely avoided (energy used to bias P1 104 and P2 106 is
repeatedly recovered) after initial charging. (The figure of merit
FOM=QG*RDS will generally be approximately constant for a given
process technology). Accordingly, the configuration shown in FIG. 1
enabling P1 104 and P2 106 to be self-driven is relatively
non-dissipative compared to gate driving using control signals
because energy is taken back from parasitic capacitances C1 114 and
C2 122. Energy "recycling" is performed by discharging capacitances
and turning voltage into current through the inductor 112.
Switching losses are thereby minimized. (See also similar
cross-coupled switches in FIG. 7.)
[0029] The sources of N1 116 and N2 124 are preferably connected
directly to GND 118, reducing the power required to bias N1 116 and
N2 124. Also, N1 116 and N2 124 can be driven using fixed
frequency, fixed or adaptive duty cycle signals. Driving N1 116 and
N2 124 at a fixed frequency, with a fixed or adaptive duty cycle,
reduces the complexity and power cost of controlling the MOSFETS.
(Duty cycle can be adaptive to maintain zero voltage switching and
zero current switching conditions despite variations of circuit
parameters or switching frequency, as further described with
respect to FIG. 2.)
[0030] Preferably, RMS currents are minimized to thereby lower
conduction losses.
[0031] FIG. 2 shows a transistor-level diagram of an isolated DC-DC
converter 200. The converter 200 can be broadly described as two
instances of the circuit portion 100, one on the input side 202 and
one on the output side 204, except that the input side 202 also
comprises a capacitor connected between Vin 206 and input ground
GND_P 208 acting as an input filter 210, and the output side 204
also comprises a capacitor connected between Vout 212 and output
ground GND_S 214 acting as an output filter 216 (input and output
filters 210, 216 are further described hereinbelow). A transformer
218 comprises a primary side 220 and a secondary side 222. The
primary side 220 and secondary side 222 of the transformer 218 each
preferably function similarly, with respect to their respective
side of the converter 200 (i.e., the input side 202 and the output
side 204), to the inductor 112 of the circuit portion 100 described
with respect to FIG. 1.
[0032] As shown in FIG. 2, on the input side 202, input voltage Vin
206 is connected to the sources of two PMOS transistors, P1P 224
and P2P 226. Nodes V1P 228 and V2P 230 are cross-coupled between
the drains and gates of P1P 224 and P2P 226. Accordingly, node V1P
228 is connected to the drain of P1P 224 and the gate of P2P 226;
and node V2P 230 is connected to the drain of P2P 226 and the gate
of P1P 224. The primary side 220 of the transformer 218 is
connected between V1P 228 and V2P 230. V1P 228 is also connected to
the drain of NMOS transistor N1P 232, and V2P 228 is also connected
to the drain of NMOS transistor N2P 234.
[0033] Control signal G1P 236 (switch control voltage) is connected
(through a buffer) to the gate of N1P 232, and control signal G2P
238 is connected (through a buffer) to the gate of N2P 234. The
sources of both N1P 232 and N2P 234 are connected to primary side
ground GND_P 208. (Parasitic capacitances corresponding to C1 114
and C2 122 exist, but are not shown.)
[0034] On the output side 204, Vout 212 is connected to the sources
of two PMOS transistors, P1S 240 and P2S 242. Nodes V1S 244 and V2S
246 are cross-coupled between the drains and gates of P1S 240 and
P2S 242. Accordingly, node V1S 244 is connected to the drain of P1S
240 and is connected to the gate of P2S 242; and node V2S 246 is
connected to the drain of P2S 242 and is connected to the gate of
P1S 240. The secondary side 204 of the transformer 218 is connected
between V1S 244 and V2S 246. V1S 244 is also connected to the drain
of NMOS transistor N1S 248, and V2S 246 is also connected to the
drain of NMOS transistor N2S 250.
[0035] Control signal G1S 252 is connected (through a buffer) to
the gate of N1S 248, and control signal G2S 254 is connected
(through a buffer) to the gate of N2S 250. The sources of both N1S
248 and N2S 250 are connected to secondary side ground GND_S 214.
(Parasitic capacitances corresponding to C1 114 and C2 122 exist,
but are not shown.)
[0036] Vin 206 is preferably connected to GND_P 208 through a
capacitor acting as an input filter 210 (a decoupling capacitor),
which reduces the "ripple"--the noise--in the power signal
transmitted from the primary side 220 to the secondary side 222 of
the transformer 218. The input filter 210 also acts as a local
bypass element to keep on-chip supply voltage stable regardless of
voltage noise generated by the input current, rapid parametric
variations in the package, and bond-wires series resistance and
inductance. Similarly, Vout 212 is preferably connected to GND_S
214 through a decoupling capacitor acting as an output filter 216,
which reduces the ripple in the output of the converter 200, and
acts as a local bypass element to keep on-chip supply voltage
stable, similarly to the input filter 210.
[0037] Preferably, in embodiments as shown in FIG. 2, the voltage
difference between Vin 206 and GND_P 208, and the voltage
difference between Vout 212 and GND_S 214, are both about 5 volts
or less. This enables use of smaller (smaller device area) input
and output filters 210, 216 that can be fabricated using relatively
inexpensive and/or straightforward processing techniques
(preferably the same class of techniques as used for the majority
of other converter components), e.g., CMOS processing. Accordingly,
capacitors rated for approximately 5 volts or less are generally
higher density (take up less device area), cheaper to fabricate and
can be fabricated with fewer additional process steps than
capacitors rated for significantly more than 5 volts (e.g., 10 or
15 volts).
[0038] Switching on the input side 202 and on the output side 204
is preferably adaptive. Accordingly, current through the primary
and secondary sides 220, 222 of the transformer 218 is
(respectively) tracked at NMOS transistor turn-on on the primary
side 220, and at NMOS transistor turn-off on the secondary side
222, to check whether current is positive or negative. Switch
timing--signal timing for G1P 236, G2P 238, G1S 252 and G2S 254--is
then adjusted to enable zero current switching and zero voltage
switching of respective NMOS transistors.
[0039] G1P 236 and G2P 238 are preferably controlled to switch on
N1P 232 and N2P 234, respectively, when zero voltage is across
switch N1P 232 or N2P 234, respectively. Preferably, turn-off of
primary side switches N1P 232 and N2P 234 is controlled by a fixed
timing (fixed frequency and adaptive duty cycle) selected to obtain
zero voltage turn-on of N2P 234 and N1P 232 (respectively) in the
subsequent phase of the cycle. Accordingly, the transition time
between switching off N1P 232 and N2P 234 is selected to enable the
transformer current to charge V1P 228 and V2P 230 to Vin (in
respective cycle phases) and subsequently discharge V2P 230 and V1P
228 (respectively) to GND_P 208, in order to turn on N2P 234 and
N1P 232 (respectively) at zero voltage across the switches N2P 234
and N1P 232.
[0040] G1S 252 and G2S 254 are preferably controlled to switch off
N1S 248 and N2S 250, respectively, when zero current flows through
the secondary side 222 of the transformer 218, and zero voltage is
across switches N1S 248 and N2S 250. Preferably, turn-on of
secondary side switches N1S 248 and N2S 250 is controlled by a
fixed timing (fixed frequency and adaptive duty cycle) according to
a selected switching frequency of the converter 200.
[0041] Input side 202 and output side 204 switch timings are also
preferably independent of each other. Preferably, output side
timings are similar to (or the same as) input side timings in
having the same switching frequency and duty cycle, but with a
phase delay with respect to input side timings--accordingly, output
side timings are phase shifted later than corresponding input side
timings.
[0042] In some embodiments, G1S 252 and G2S 254 can be tied to
GND_S 214 or otherwise connected such that N S 248 and N2S 250
remain always off (providing a diode-like effect) and perform
passive rectification (as described further with respect to FIGS. 3
and 5). This can generally be done without compromising function of
the output side 204, except that conduction losses from N1S 248 and
N2S 250 will be higher than when they are switched (switching of
N1S 248 and N2S 250 is further described with respect to FIG. 3).
The increase in conduction losses from not switching N1S 248 and
N2S 250 can be less relevant in high voltage applications, and may
come with efficiency gains from avoiding use of control circuits to
generate G1S 252 and G2S 254 signals. Whether switching N1S 248 and
N2S 250 results in improved overall performance for the converter
200--as determined with respect to efficiency and/or other
metric(s)--may be application-dependent.
[0043] FIG. 3 shows a timing diagram 300 for an isolated DC-DC
converter 200 as shown in and described with respect to FIG. 2.
[0044] In FIG. 3, VLP 302 is the voltage across the primary side
220 of the transformer 218 and equals the voltage at V1P 228 minus
the voltage at V2P 230. Voltage across the primary side 220 is
therefore positive when V1P 228 is higher voltage than V2P 230 and
negative when V2P 230 is higher voltage than V1P 228. ILP 304 is
the current across the primary side 220 from node V1P 228 to node
V2P 230, and is positive flowing from V1P 228 to V2P 230 and
negative flowing from V2P 230 to V1P 228.
[0045] VLS 306 is the voltage across the secondary side 222 of the
transformer 218 and equals the voltage at V1S 244 minus the voltage
at V2S 246. Voltage across the secondary side 222 is therefore
positive when V1S 244 is higher voltage than V2S 246 and negative
when V2S 246 is higher voltage than V1S 244. ILS 308 is the current
across the secondary side 222 from node V1S 244 to node V2S 246,
and is positive flowing from V1S 244 to V2S 246 and negative
flowing from V2S 246 to V1S 244.
[0046] As shown in the timing diagram 300 of FIG. 3, at time
T.sub.0 G1P 236 is low, meaning N1P 232 is off, and G2P 238 is
high, meaning N2P 234 is on. Because N2P 234 is on, V2P 230 and the
gate of P1P 224 are connected to GND_P 208 and are therefore pulled
low. As a result, P1P 224 is on, connecting V1P 228 and the gate of
P2P 226 to Vin 206 and thereby pulling V1P 228 high, which causes
P2P 226 to be off. This means that VLP 302 is high (V1P 228 is high
and V2P 230 is low). ILP 304 (current across the primary side 220)
has a slope determined by the interaction of Vin 206, Vout 212, and
transformer parameters (e.g., the transformer turns ratio and the
leakage inductance of the transformer 218 on the primary side 220).
VLP 302 being high will contribute to increasing the slope of ILP
304 (in a positive direction).
[0047] At time T.sub.1, G2P 238 goes low, switching off N2P 234 so
that GND_P 208 is no longer pulling down V2P 230. N2P 234 is turned
off at a time selected to enable V2P 230 to be fully charged and
V1P 228 to be fully discharged by the current through the inductor,
ILP 304, before G1P 236 going high to switch on N1P 232 (i.e., to
enable zero voltage turn-on of N1P 232). The current across the
primary side 220, ILP 304, subsequently drives charge into the
parasitic capacitances connected to V2P 230 (corresponding to C2
122, see FIG. 1)--which includes the gate-source capacitance of P1P
224, as described hereinabove with respect to FIG. 1--causing V2P
230 to increase and causing VLP 302 to fall to zero.
[0048] At time T.sub.2, V2P 230 is high (fully charged by the
current across the inductor 218, as described with respect to FIG.
2) and VLP 302 reaches zero voltage--i.e., zero voltage difference
exists between V1P 228, which is high, and V2P 230, which is also
high. V2P 230 being high causes P1P 224 to turn off, so that V1P
228 is no longer pulled up by Vin 206. This allows ILP 304 to drive
charge out of P2P's 226 gate capacitance and other parasitic
capacitances at node V1P 228 (corresponding to C1 114, see FIG. 1)
and recycle that charge as current across the primary side 220,
resulting in V1P 228 going low, which causes VLP 302 to continue to
fall. VLP 302 below zero contributes to decreasing the slope of ILP
304 (increasing the slope of ILP 304 in a negative direction).
[0049] At time T.sub.3 316, V1P 228 reaches its low state, so G1P
236 is controlled to go high to turn on N1P 232 (zero voltage
switch on, i.e., ZVS). This connects V1P 228 to GND_P 208, pulling
V1P 228 low. V1P 228 is already low when connected to GND_P 208, so
no (or minimal) losses occur from discharge of parasitic
capacitances connected to V1P 228 (corresponding to C1 114). V1P
228 being low turns on P2P 226, which connects V2P 230 to Vin 206,
pulling V2P 230 high.
[0050] As shown in the timing diagram 300, signal behaviors of the
converter 200 at times T.sub.4, T.sub.5 and T.sub.6 echo the signal
behaviors described with respect to FIG. 3 at times T.sub.1,
T.sub.2 and T.sub.3, respectively, except: voltages and currents
across the primary side 220 have the opposite sign; and P2P 226,
V2P 230, N2P 234 and G2P 238 have the behaviors described
hereinabove with respect to P1P 224, V1P 228, N1P 232 and G1P 236,
and vice versa. Accordingly, T.sub.1, T.sub.2, T.sub.3 cover a half
cycle of operation of the converter 200, and T.sub.4, T.sub.5 and
T.sub.6 cover another half cycle of operation of the converter 200
(in the same cycle). Operation of the converter 200 (switch
behaviors) during the half cycle covered by T.sub.4, T.sub.5 and
T.sub.6 is the same as operation of the converter 200 during the
half cycle covered by T.sub.1, T.sub.2, T.sub.3, except that the
roles of switches (on both the input side 202 and the output side
204) on different sides of respective inductors (the primary side
220 and the secondary side 222) are exchanged to enable current to
flow in the reverse direction. (This can also be thought of as
repeating switch and signal behaviors of times T.sub.1, T.sub.2 and
T.sub.3 at times T.sub.4, T.sub.5 and T.sub.6 as if the terminals
of the primary side 220 are reversed, and the terminals of the
secondary side 222 are reversed.)
[0051] Example timings for output side 204 switching and signals
can be determined from FIG. 3. The output side 204 timings shown in
FIG. 3 are examples, as described hereinabove, and output side 204
timings are generally independent from input side 202 timings.
[0052] For example, output side 204 switch control signals G1S 252
and G2S 254 will preferably turn on NMOS transistors N1S 248 and
N2S 250 when voltage across the respective transistor is zero (at a
time corresponding to a specified switching frequency; preferably,
as soon as the voltage across the transistor reaches zero to
maximize power transmitted to Vout 212). This also corresponds to
voltage across the secondary side 222, VLS 306, being at a maximum
positive or negative value, meaning that one of V1S 244 or V2S 246
is low, and the other is high. Accordingly, if V1S 244 is low
(voltage across N1S 248 is zero), then G1S 252 switches on N1S 248,
and if V2S 246 is low (voltage across N2S 250 is zero), then G2S
254 switches on N2S 250 (zero voltage turn-on). Also, G1S 252 or
G2S 254 will switch off NMOS transistor N1S 248 or N2S 250
(respectively) when current through the secondary side 222, ILS
308, is zero (zero current turn-off) and voltage across the
respective NMOS transistor being turned off is zero (zero voltage
turn-off).
[0053] As shown in FIG. 5, in some embodiments, gates of low-side
NMOS transistors N1S 248 and N2S 250 can be connected to ground so
that they perform passive rectification on currents through the
secondary side (ILS 308). This prevents negative output current
(from Vout 212 to GND_S 214) and can reduce the need for complex
rectification circuitry, at the cost of the voltage drop across N1S
248 and N2S 250 diodes.
[0054] Active (or synchronous) rectification is performed by
low-side NMOS transistors N1P 232, N2P 234, N1S 248 and N2S 250
switching on to conduct current when the current is or is about to
flow in the correct (positive) direction (from Vin 206 to GND_P
208, or from GND_S 214 to Vout 212), thereby preventing (negative)
current in the opposite direction. Thus, when V1P 228, V2P 230, V1S
244 or V2S 246 might otherwise be too high, potentially causing
power losses by resulting in negative current or (on the output
side 104) reduced output voltage (at Vout 212), the problematic
node will be grounded by N1P 232, N2P 234, N1S 248 or N2S 250,
respectively. Switch and signal timings described with respect to,
e.g., FIGS. 2 and 3 are used to achieve these results. Secondary
side diodes (of N1S 248 and N2S 250) also help to prevent negative
output current; accordingly, in the converter 200 as shown in FIG.
2, passive rectification is performed alongside active
rectification.
[0055] FIG. 4 shows a transistor-level diagram of a high output
voltage isolated DC-DC converter 400. In embodiments as shown in
FIG. 4, the input side 402 of the converter 400 is arranged the
same as the input side 202 of the converter 200 (see FIG. 2). The
output side 404 of FIG. 4 comprises two instances of the output
side 204 of the converter 200, an output instance 406 and a ground
instance 408. The output instance 406 and the ground instance 408
are "stacked". Accordingly, Vout 212 of the output instance 406
comprises Vout 410 of the converter 400; GND_S 214 of the ground
instance 408 comprises GND_S 412 of the converter 400; and GND_S
214 of the output instance 406 is connected to Vout 212 of the
ground instance 408 at Vout/2 414.
[0056] Also, V1S 244 and V2S 246 are connected to the terminals of
the secondary side 222 of the transformer 218 through isolation
capacitors 416.sub.C1 and 416.sub.C2 which isolate the respective
V1S 244 and V2S 246 nodes of the output instance 406 and the ground
instance 408 from each other. Preferably, an output capacitor
418.sub.O is connected between Vout 410 and Vout/2 414, and another
output capacitor 418.sub.G is connected between Vout/2 414 and
GND_S 412.
[0057] The stacked output instance 406 and ground instance 408 act
as a voltage doubler, with the input voltage of Vin 206 as the base
(doubled) voltage. Vout/2 414 acts as the voltage output for the
ground instance 408 and as the ground for the output instance 406
(the ground instance and output instance are otherwise isolated
from each other by the isolation capacitors 416.sub.C1 and
416.sub.C2). Consequently, the voltage at Vout/2 414 is one half of
the voltage at Vout 410 (and this voltage is also available as an
output). The ground instance 408 and the output instance 406 each
function similarly to the output side 204 of converter 200 of FIG.
2.
[0058] This means that the converter 400 of FIG. 4 retains the
efficiency advantages of the converter 200 of FIG. 2, while being
able to output twice as high a voltage--e.g., 10 volts output given
a 5 volt input--without requiring components rated for a
correspondingly higher voltage. For example, for a 5 volt input and
10 volt output, preferably, isolation capacitors 416 and output
capacitors 418.sub.O, 418.sub.G are 5 volt capacitors; and the
various NMOS and PMOS transistors on the output side 204 of the
converter 400 can be 5 volt transistors; with capacitors and
transistors preferably fabricated using relatively inexpensive
and/or straightforward processing techniques, preferably the same
class of techniques, e.g., CMOS processing. Enabling high voltage
output without requiring high voltage components lowers process
complexity (e.g., component complexity, number of process steps and
number of masks required) and therefore process cost and required
device area, and preserves applicability of CMOS fabrication (which
is relatively simple and low cost and avoids fabrication process
limitations). This approach also enables use of a relatively low
transformer turns ratio, e.g., 1:1, maintaining transformer
symmetry.
[0059] FIG. 5 shows a transistor-level diagram of a high output
voltage isolated DC-DC converter 500 in which the bias of the NMOS
transistors in the ground instance 408 of the output side 404, N1S
248 and N2S 250, are respectively connected to their own source, so
that transistors N1S 248 and N2S 250 (of the ground instance 408)
act as diodes. In embodiments as shown in FIG. 5, the resonance
(the secondary side 214) is sufficient to drive charge into and out
of the respective nodes V1S 242 and V2S 244 of the output instance
406 and the ground instance 408 to implement self-switching of
respective PMOS transistors P1S 240 and P2S 242 and thereby drive
the secondary side 222. This avoids switching losses from N1S 248
and N2S 250 without using control signals and without paying the
power (and device area) cost of corresponding control
circuitry.
[0060] However, the voltage drop across the NMOS transistors N1S
248 and N2S 250 in the ground instance 408 of the converter 500
shown in and described with respect to FIG. 5 is slightly higher
than the voltage drop across the NMOS transistors N1S 248 and N2S
250 (switched using control signals G1S 252 and G2S 254) in the
converter 400 shown in and described with respect to FIG. 4. This
is because N1S 248 and N2S 250 in the ground instance 408 of the
converter 500 in FIG. 5 are configured to act as diodes, rather
than as switched transistors. As a result, slightly higher
conduction losses occur in a converter 500 of FIG. 5 than in a
converter 400 of FIG. 4. (In the ground instance 408 of the
converter 500 of FIG. 5, N1S 248 and N2S 250 prevent negative
output current due to their function as diodes.)
[0061] It may therefore be application-dependent as to whether a
converter 400 (as in FIG. 4) or a converter 500 (as in FIG. 5) is
more suited to a particular purpose or otherwise performs better
with respect to application-related performance metrics.
[0062] FIG. 6A shows an example of a process 600 for controlling an
isolated DC-DC converter. As shown in FIG. 6A, an input side 202
low side switch (e.g., NMOS) is turned on when zero voltage is
across the low side switch, connecting a corresponding first
primary side 220 transformer terminal to the primary side ground
208 (GND_P 208) in step 602. Zero voltage at the first primary side
220 terminal causes a high side switch (e.g., a PMOS that has its
gate coupled to the first terminal, its drain coupled to a second
primary side 220 transformer terminal and its source coupled to Vin
206) to turn on, connecting the second terminal to Vin 206 in step
604. After a period selected to enable a current through the
primary side 220 to (after the low side switch is turned off)
charge the first terminal to an input voltage 206 and discharge the
second terminal to a primary side ground 208 voltage, the low side
switch is turned off, disconnecting the first terminal from ground
208 in step 606. As a result, the gate of the self-driven high side
switch is released (from ground 208), which allows the high side
switch gate to be driven by current through the primary side 220 to
input voltage Vin 206; this causes the high side switch to open,
disconnecting the second terminal from Vin 206 in step 608. The
current through the primary side 220 then drives the voltage at the
second terminal to ground (GND_P 208) in step 610. The process is
then repeated in the opposite direction (as if the terminals were
reversed, i.e., the first terminal substituted for the second and
vice versa, using different corresponding high side and low side
switches) in step 612.
[0063] The disoverlap time between the turn off of the first low
side switch and the turn on time of the second low side switch is
selected to allow the full voltage swing of the switching nodes,
enabling zero voltage switching of the second low side switch.
[0064] FIG. 6B shows an example of a process 614 for controlling an
isolated DC-DC converter. As shown in FIG. 6B, in step 616, an
output side 204 low side (e.g., NMOS) switch is turned on at a
specified time (corresponding to a specified switching frequency)
when zero voltage is across the low side switch; turning on the low
side switch connects a corresponding first secondary side 222
transformer terminal to the secondary side ground 214 (GND_S 214).
The resulting zero voltage at the first secondary side 222 terminal
causes a high side switch (e.g., PMOS) which is coupled to (when
activated) connect a second secondary side 222 transformer terminal
to Vout 212 to turn on, connecting the second terminal to Vout 212
in step 618. If zero current flows through the secondary side 222,
and zero voltage is across the low side switch, then the low side
switch is turned off, disconnecting the first secondary side 222
terminal from the secondary side ground 214 in step 620. This
allows the first secondary side 222 terminal, and thus the high
side switch gate, to be driven to output voltage Vout 212 by
current through the secondary side 222. As a result, the gate of
the high side switch opens, disconnecting the second secondary side
222 terminal from Vout 212 in step 622. The current through the
secondary side 222 then drives the voltage at the second secondary
side 222 terminal to the secondary side ground voltage in step 624.
The process is then repeated in the opposite direction (with
different corresponding high side and low side switches) in step
626.
[0065] FIG. 7 shows an example of a transistor-level diagram of a
circuit portion 700 used to construct an isolated DC-DC converter.
As shown in FIG. 7, the drains of two NMOS transistors HS1 702
(high side 1) and HS2 704 (high side 2) are connected to a voltage
input Vin 706. The gate of HS1 702 is connected (via a buffer) to a
control voltage G1 708, and the gate of HS2 704 is connected (via a
buffer) to a control voltage G2 710. The source of HS1 702 is
connected to node V1 712, and the source of HS2 704 is connected to
node V2 714.
[0066] V1 712 is connected to a first terminal of an inductor 716,
and V2 714 is connected to a second terminal of the inductor 716.
An NMOS transistor LS1 718 (low side 1) is connected between V1 712
and a ground GND 720. An NMOS transistor LS2 722 (low side 2) is
connected between V2 714 and GND 720. The sources of LS1 718 and
LS2 722 are connected to GND 720. The gates and drains of LS1 718
and LS2 722 are cross-coupled between V1 712 and V2 714.
Accordingly, the drain of LS1 718 is connected to V1 712, and the
gate of LS1 718 is connected to V2 714 through a voltage limiter
VL1 724; and the drain of LS2 722 is connected to V2 714, and the
gate of LS2 722 is connected to V1 712 through a voltage limiter
VL2 726. The bias voltage of LS1 718 is V2 714, limited by voltage
limiter VL1 724, and the bias voltage of LS2 722 is V1 712, limited
by voltage limiter VL2 726. Voltage limiters VL1 724 and VL2 726
are further described with respect to FIG. 10.
[0067] The inductor 716 can also be considered a primary side or a
secondary side of a transformer--accordingly, as described with
respect to FIG. 1, the primary side or secondary side of a
transformer is an inductor. Circuit portion 700 can be used to
construct an isolated DC-DC converter such that the primary side or
secondary side of a transformer in the converter stands in place of
the inductor 716, as shown with respect to FIGS. 8 and 10.
[0068] In operation, the NMOS transistors HS1 702 and HS2 704 are
preferably zero voltage switched and zero current switched.
Accordingly, G1 708 and G2 710 are controlled to turn on OR off
(generally one or the other and not both; the other of turn-on or
turn-off timing is determined using other criteria, such as further
described with respect to FIG. 9) HS1 702 and HS2 704 when
corresponding adjacent source nodes V1 712 and V2 714 are fully
charged to Vin 706, and zero current flows across the inductor 716
(from node V1 712 to node V2 714 or vice versa). (Zero voltage
switching and zero current switching are further described with
respect to FIGS. 3 and 9.) This results in minimal switching
losses, because little or no voltage is across (or little or no
current is conducted by) HS1 702 or HS2 704 (for either the turn-on
or turn-off transition), while they are switched.
[0069] In embodiments as shown in FIG. 7, power cost for MOSFET
control is reduced by making LS1 718 and LS2 720 self-driving.
Accordingly, rather than requiring control signals to bias LS1 718
and LS2 720, LS1 718 and LS2 720 are biased by charge driven to LS1
718 and LS2 720 via nodes V2 714 and V1 712 (respectively; and
through voltage limiters VL1 724 and VL2 726, respectively) by the
current through the resonance (inductor 716). Self-driving LS1 718
and LS2 720 reduces the complexity and power cost of controlling
the MOSFETS.
[0070] Further power savings are achieved by using the current
across the inductor 716 (the resonance current) to drive charge out
of the gates of LS1 718 and LS2 720 and turn that charge into
current across the inductor 716. Some of that current charges
respective drain-source parasitic capacitances of adjacent high
side and low side transistors, and gate-source parasitic
capacitances of low side transistors. This recycling avoids
dissipation of gate charges (Cgs charges) when LS1 718 and LS2 720
are switched, because those gate charges are reused. Accordingly,
those gate charges (charges of gate-source parasitic capacitances),
along with charges of respective drain-source parasitic
capacitances, are driven out of parasitic capacitances adjacent to
V1 712 or V2 714 by the resonance current, turned into current
across the inductor 716, and (in part) are used to charge parasitic
capacitances adjacent to V2 714 or V1 712 (respectively).
[0071] Because gate charge of LS1 718 and LS2 720 are reused, the
tradeoff between QG (gate charge) and RDS (drain-source on
resistance) can be largely avoided for LS1 718 and LS2 720. For
example, a lower RDS can generally be used, thereby enabling faster
switching and lower heat output, without resulting in excessive
power costs due to a higher QG, because the gate charge power cost
is largely avoided (energy used to bias LS1 718 and LS2 720 is
repeatedly recovered) after initial charging. (The figure of merit
FOM=QG*RDS will generally be approximately constant for a given
process technology). Accordingly, the configuration shown in FIG. 7
enabling LS1 718 and LS2 720 to be self-driven is relatively
non-dissipative compared to gate driving using control signals
because energy is taken back from parasitic capacitances. Energy
"recycling" is performed by discharging capacitances and turning
voltage into current through the inductor 716. Switching losses are
thereby minimized.
[0072] The sources of LS1 718 and LS2 720 are preferably connected
directly to GND 720, reducing the power required to bias LS1 718
and LS2 720. Also, LS1 718 and LS2 720 can be driven using fixed
frequency, fixed or adaptive duty cycle signals. Driving LS1 718
and LS2 720 at a fixed frequency, with a fixed or adaptive duty
cycle, reduces the complexity and power cost of controlling the
MOSFETS. (Duty cycle can be adaptive to maintain zero voltage
switching and zero current switching conditions despite variations
of circuit parameters or switching frequency.)
[0073] Preferably, RMS currents are minimized to thereby lower
conduction losses.
[0074] FIG. 8 shows a transistor-level diagram of an isolated DC-DC
converter 800. The converter 800 can be broadly described as two
instances of the circuit portion 700, one on the input side 802 and
one on the output side 804. A transformer 806 comprises a primary
side 808 and a secondary side 810. The primary side 808 and
secondary side 810 of the transformer 806 each preferably function
similarly, with respect to their respective side of the converter
800 (i.e., the input side 802 and the output side 804), to the
inductor 716 of the circuit portion 700 described with respect to
FIG. 7.
[0075] As shown in FIG. 8, on the input (i.e., primary) side 802,
input voltage Vin 812 is connected to the drains of two high side
NMOS transistors, HS1P 814 and HS2P 816. HS1P 814 and HS2P 816 are
switched by control signals G1P 818 and G2P 820 connected (through
buffers) to the gates of HS1P 814 and HS2P 816, respectively. Node
V1P 822 is connected to the source of HS1P 814, and node V2P 824 is
connected to the source of HS2P 816.
[0076] The drains and gates of low side NMOS transistors LS1P 826
and LS2P 828 are cross-coupled between V1P 822 and V2P 824, the
connections between nodes (V2P 824 and V1P 822) and gates (of LS1P
826 and LS1P 828) being made through voltage limiters VL1P 830 and
VL2P 832, respectively. Accordingly, the drain of LS1P 826 is
connected to V1P 822, and the gate of LS1P 826 is connected to VL1P
830 which is connected to V2P 824; and the drain of LS2P 828 is
connected to V2P 824, and the gate of LS2P 828 is connected to VL2P
832 which is connected to V1P 822. Voltage limiters VL1P 830 and
VL2P 832 limit the voltage that reaches the gates of LS1P 826 and
LS1P 828, respectively. For example, if Vin 812 is 12 volts, and
LS1P 826 and LS1P 828 are fabricated to be biased by 5 volts, then
VL1P 830 and VL2P 832 limit the voltages that reach the gates of
LS1P 826 and LS1P 828, respectively, to 5 volts.
[0077] The sources of LS1P 826 and LS2P 828 are connected to
primary side ground GND_P 834. The primary side 808 of the
transformer 806 is connected between V1P 822 and V2P 824.
[0078] On the output side 804, output voltage Vout 834 is connected
to the drains of two high side NMOS transistors, HS1S 836 and HS2S
838. HS1S 836 and HS2S 838 are switched by control signals G1S 840
and G2S 842 connected (through buffers) to HS1S 836 and HS2S 838,
respectively. Node V1S 844 is connected to the source of HS1S 836,
and node V2S 846 is connected to the source of HS2S 838.
[0079] The drains and gates of NMOS transistors LS1S 848 and LS2S
850 are cross-coupled between V1S 844 and V2S 846, the connections
between nodes (V2S 846 and V1S 844) and gates (of LSS 848 and LSS
850) being made through voltage limiters VLS 852 and VL2S 854,
respectively. Accordingly, the drain of LS1S 848 is connected to
V1S 844, and the gate of LSS 848 is connected to VL1S 852 which is
connected to V2S 846; and the drain of LS2S 850 is connected to V2S
846, and the gate of LS2S 850 is connected to VL2S 854 which is
connected to V1S 844. The sources of LSS 848 and LS2S 850 are
connected to secondary side ground GND_S 856.
[0080] The secondary side 810 of the transformer 806 is connected
between V1S 844 and V2S 846.
[0081] Switching on the input side 802 and on the output side 804
is preferably adaptive. Accordingly, current through the primary
and secondary sides 808, 810 of the transformer 806 is
(respectively) tracked at high side transistor turn-on on the
primary side 808, and at high side transistor turn-off on the
secondary side 810, to check whether current is positive or
negative. Switch timing--signal timing for G1P 818, G2P 820, G1S
840 and G2S 842--is then adjusted to enable zero current switching
of respective high side transistors. G1P 818 and G2P 820 are
preferably controlled to switch on HS1P 814 and HS2P 816,
respectively, when zero current flows through the primary side 808
of the transformer 806 and zero voltage is across the transistor
being switched on. G1P 818 and G2P 820 are preferably controlled to
switch off HS1P 814 and HS2P 816, respectively, at a time selected
to enable the current across the primary side to (after the high
side transistor is switched off) fully discharge V1P 822 or V2P 824
(respectively) to the primary side ground voltage and fully charge
V2P 824 or V1P 822 (respectively) to the input voltage. G1S 840 and
G2S 842 are preferably controlled to switch on HS1S 836 and HS2S
838, respectively, when zero voltage is across the transistor being
switched on, at a time selected to correspond to a switching
frequency dependent on the switching frequency of the primary side,
and dependent on capacitor and inductor parameters. G1S 840 and G2S
842 are preferably controlled to switch off HS1S 836 and HS2S 838,
respectively, when zero current flows through the secondary side
810 of the transformer 806 and zero voltage is across the
transistor being switched off. Input side 802 and output side 804
switch timings are also preferably independent of each other.
[0082] The converter 800 also preferably comprises an input
decoupling capacitor 858 connected between Vin 812 and GND_P 832,
and an output decoupling capacitor 860 connected between Vout 834
and GND_S 856 (as described hereinabove for decoupling capacitors,
also called ripple capacitors, of FIG. 2).
[0083] FIG. 9 shows an example of a timing diagram 900 for an
isolated DC-DC converter 800 as shown in FIG. 8. In FIG. 9, VLP 902
is the voltage across the primary side 808 of the transformer 806
and equals the voltage at V1P 822 minus the voltage at V2P 824.
Voltage across the primary side 808 is therefore positive when V1P
822 is higher voltage than V2P 824 and negative when V2P 824 is
higher voltage than V1P 822. ILP 904 is the current across the
primary side 808 from node V1P 822 to node V2P 824, and is positive
flowing from V1P 822 to V2P 824 and negative flowing from V2P 824
to V1P 822.
[0084] VLS 906 is the voltage across the secondary side 810 of the
transformer 806 and equals the voltage at V1S 844 minus the voltage
at V2S 846. Voltage across the secondary side 810 is therefore
positive when V1S 844 is higher voltage than V2S 846 and negative
when V2S 846 is higher voltage than V1S 844. ILS 908 is the current
across the secondary side 810 from node V1S 844 to node V2S 846,
and is positive flowing from V1S 844 to V2S 846 and negative
flowing from V2S 846 to V1S 844.
[0085] At time T.sub.0, V1P 822 is low, V2P 824 is high, G1P 818 is
low, and G2P 820 is high. Because G1P 818 is low, HS1P 814 is off.
Because G2P 820 is high, HS2P 816 is on, connecting V2P 824 to Vin
812, pulling V2P 824 high. Because V1P 822 is low, LS2P 828 is off.
Because V2P 824 is high, LS1P 826 is on, connecting V1P 822 to GND
834, pulling V1P 822 low. Also, VLP 902 is negative, and ILP 904
has a slope determined by Vin 812, Vout 834 and parameters of the
transformer 806 (as described hereinabove for current with respect
to FIG. 3). At time T.sub.0, the voltage across HS2P 816 is zero,
because the source-drain path of HS2P 816 is connected to Vin 812
and V2P 824, and V2P 824 is high.
[0086] At time T.sub.1, while the voltage across HS2P 816 is zero
and the current through the inductor ILP 904 is zero, G2P 820 is
controlled to go low, turning off HS2P 816 and thereby achieving
the desired zero voltage and zero current turn-off (ZVS and ZCS).
As a result, charge in parasitic capacitances connected to V2P 824
is converted into current across the primary side 808 (ILP 904),
causing V2P 824 (and therefore VLP 902) to fall.
[0087] At time T.sub.2, V2P 824 is zero, turning off LS1P 826. VLP
902 is also zero (V1P 822 and V2P 824 are both zero). As a result,
current across the primary side 808 (ILP 904) charges the parasitic
capacitances connected to V1P 822, causing V1P 822 to rise and VLP
902 to increase above zero.
[0088] At time T.sub.3, V1P 822 is high, turning on LS2P 828, and
G1P 818 goes high, turning on HS1P 814, which pulls up V1P 822. At
time T.sub.3, the voltage across HS1P 814 is zero, because the
source-drain path of HS1P 816 is connected to Vin 812 and V1P 822,
and V2P 822 is high. Also at time T.sub.3, VLP 902 is positive.
[0089] As shown in the timing diagram 900, signal behaviors of the
converter 800 at times T.sub.4, T.sub.5 and T.sub.6 echo the signal
behaviors described with respect to FIG. 9 at times T.sub.1,
T.sub.2 and T.sub.3, respectively, except: voltages and currents
across the primary side 808 have the opposite sign; and HS1P 814,
V1P 822, LS1P 826 and G1P 818 have the behaviors described
hereinabove with respect to HS2P 816, V2P 824, LS2P 828 and G2P
820, and vice versa. (Accordingly, the switch and signal behaviors
of times times T.sub.1, T.sub.2 and T.sub.3 are preferably repeated
at times T.sub.4, T.sub.5 and T.sub.6 as if the terminals of the
primary side 808 and secondary side 810 had been reversed.)
[0090] Example timings for output side 804 switching and signals
can be determined from FIG. 9. The output side 804 timings shown in
FIG. 9 (voltage across the secondary side 810, VLS 906, and current
through the secondary side 810, ILS 908) are examples, as described
hereinabove, and output side 804 timings are generally independent
from input side 802 timings.
[0091] For example, output side 804 switch control signals G1S 840
and G2S 842 will preferably turn on HS1S 836 and HS2S 838 when VLS
906 is at a maximum positive or negative value, meaning that one of
V1S 844 or V2S 846 is low, and the other is high. Accordingly, if
V1S 844 is high, then G1S 840 switches on HS1S 836, and if V2S 846
is high, then G2S 842 switches on HS2S 838. Also, G1S 840 or G2S
842 will switch off NMOS transistor HS1S 836 or HS2S 838
(respectively) when current through the secondary side, ILS 810, is
zero (zero current turn-off).
[0092] As shown in FIGS. 8 and 9, NMOS transistors HS1P 814, HS2P
816 perform active rectification on currents across the primary
side 808 (ILP 904), and NMOS transistors HS1S 836 and HS2S 838
perform passive and active rectification on currents across the
secondary side 810 (ILS 908), as described hereinabove regarding
active and passive rectification with respect to FIG. 3. Switching
of these NMOS transistors prevents both negative input currents and
negative output currents. This can reduce the need for complex
rectification circuitry.
[0093] FIG. 10 shows an example of a transistor-level diagram of a
high voltage isolated DC-DC converter 1000. The converter 1000 of
FIG. 10 is based on the converter 800 of FIG. 8, with voltage
limiters comprising clamping NMOS transistors and bootstrap
circuits used to drive the high side transistors. Generally, the
operation and signal timings of the converter 1000 of FIG. 10
should be similar to those of FIG. 8. Vin 812 and Vout 834 can be
significantly higher than 5 volts, e.g., Vin 812 of 12 volts and
Vout 834 of 12 volts. Preferably, HS1P 814, HS2P 816, HS1S 836 and
HS2S 838 are LDMOS transistors, enabling them to function
effectively when biased by high voltages (e.g., 17 volts).
[0094] As shown in FIG. 10, starting from the converter 800 of FIG.
8: on the input side 802, voltage limiter VL1P 830 comprises an
NMOS transistor VL1P 1002 and voltage limiter VL2P 832 comprises an
NMOS transistor VL2P 1004. Both VL1P 1002 and VL2P 1004 preferably
have their gates connected to an input voltage V.sub.clampP 1006,
which acts as a clamping voltage. V.sub.clampP 1006 is preferably
set to 5 volts so that the low side transistors LS1P 826 and LS2P
828 will not be biased by more than 5 volts. This enables LS1P 826
and LS2P 828 to be fabricated as CMOS devices (usually rated for 5
volts or less).
[0095] The source of VL1P 1002 is preferably connected to the gate
of LS1P 826, a capacitor C1P 1008 which is also connected to
primary ground GND_P 834, and the source of a bootstrap
circuit-controlling NMOS transistor BS1P 1010. The drain of VL1P
1002 is preferably connected to V2P 824 and to the gate of BS1P
1010. The drain of BS1P 1010 is preferably connected to a node
V.sub.BS1P 1012, which is also connected to bootstrap capacitor
C.sub.BS1P 1014 and to supply a level shifting buffer B.sub.BS1P
1016 (bootstrap capacitor C.sub.BS1P 1014 acts as the supply for
the buffer B.sub.BS1P 1016). C.sub.BS1P 1014 is also connected to
V1P 822. Buffer B.sub.BS1P 1016 is also connected to connect G1P
818 to the gate of HS1P 814. Because the voltage at V.sub.BS1P 1012
preferably rises several volts above the maximum voltage reached by
V1P 822 (as further described hereinbelow), which itself is
preferably significantly above 5 volts (e.g., 12 volts), B.sub.BS1P
1016 is used to protect control circuits producing G1P 818 (usually
operating at 5 volts or less) from that elevated voltage, and to
enable G1P 818 to effectively control HS1P 814. An example level
shifting buffer can include multiple inverters, with successive
inverters (in the direction of travel from the lower voltage regime
to the higher voltage regime) having larger area transistors than
preceding inverters.
[0096] The source of VL2P 1004 is preferably connected to the gate
of LS2P 828, a capacitor C2P 1018 which is also connected to
primary ground GND_P 834, and the source of a bootstrap
circuit-controlling NMOS transistor BS2P 1020. The drain of VL2P
1004 is preferably connected to V1P 822 and to the gate of BS2P
1020. The drain of BS2P 1020 is preferably connected to a node
V.sub.BS2P 1022, which is also connected to bootstrap capacitor
C.sub.BS2P 1024 and to supply a level shifting buffer B.sub.BS2P
1026 (bootstrap capacitor C.sub.BS2P 1024 acts as the supply for
the buffer B.sub.BS2P 1026). C.sub.BS2P 1024 is also connected to
V2P 824. Buffer B.sub.BS2P 1026 is also connected to connect G2P
820 to the gate of HS2P 816. Because the voltage at V.sub.BS2P 1022
preferably rises several volts above the maximum voltage reached by
V2P 824 (as further described hereinbelow), which itself is
preferably significantly above 5 volts (e.g., 12 volts), B.sub.BS2P
1026 is used to protect control circuits producing G2P 820 (usually
operating at 5 volts or less) from that elevated voltage, and to
enable G2P 820 to effectively control HS2P 816.
[0097] On the output side 804, voltage limiter VL1S 852 comprises
an NMOS transistor VL1S 1028 and voltage limiter VL2S 854 comprises
an NMOS transistor VL2S 1030. Both VL1S 1028 and VL2S 1030
preferably have their gates connected to an input voltage
V.sub.clampS 1032, which acts as a clamping voltage. V.sub.clampS
1032 is preferably set to 5 volts so that the low side transistors
LS1S 848 and LS2S 850 will not be biased by more than 5 volts. This
enables LS1S 848 and LS2S 850 to be fabricated as CMOS devices
(usually rated for 5 volts or less).
[0098] The source of VL1S 1028 is preferably connected to the gate
of LS1S 848, a capacitor C1S 1034 which is also connected to
secondary ground GND_S 856, and the source of a bootstrap
circuit-controlling NMOS transistor BS1S 1036. The drain of VL1S
1028 is preferably connected to V2S 846 and to the gate of BS1S
1036. The drain of BS1S 1036 is preferably connected to a node
V.sub.BS1S 1038, which is also connected to bootstrap capacitor
C.sub.BS1S 1040 and to supply a level shifting buffer B.sub.BS1S
1042 (bootstrap capacitor C.sub.BS1S 1040 acts as the supply for
the buffer B.sub.BS1S 1042). C.sub.BS1S 1040 is also connected to
V1S 844. B.sub.BS1S 1042 is also connected to connect G1S 840 to
the gate of HS1S 836. Because the voltage at V.sub.BS1S 1038
preferably rises several volts above the maximum voltage reached by
V1S 844 (as further described hereinbelow), which itself is
preferably significantly above 5 volts (e.g., 12 volts), B.sub.BS1S
1042 is used to protect control circuits producing G1S 840 (usually
operating at 5 volts or less) from that elevated voltage, and to
enable G1S 840 to effectively control HS1S 836.
[0099] The source of VL2S 1030 is preferably connected to the gate
of LS2S 850, a capacitor C2S 1044 which is also connected to
secondary ground GND_S 856, and the source of a bootstrap
circuit-controlling NMOS transistor BS2S 1046. The drain of VL2S
1030 is preferably connected to V2S 846 and to the gate of BS2S
1046. The drain of BS2S 1046 is preferably connected to a node
V.sub.BS2S 1048, which is also connected to bootstrap capacitor
C.sub.BS2S 1050 and to supply a level shifting buffer B.sub.BS2S
1052 (bootstrap capacitor C.sub.BS2S 1050 acts as the supply for
the buffer B.sub.BS2S 1052). C.sub.BS2S 1050 is also connected to
V2S 846. B.sub.BS2S 1052 is also connected to connect G2S 842 to
the gate of HS2S 838. Because the voltage at V.sub.BS2S 1048
preferably rises several volts above the maximum voltage reached by
V2S 846 (as further described hereinbelow), which itself is
preferably significantly above 5 volts (e.g., 12 volts), B.sub.BS2S
1052 is used to protect control circuits producing G2S 842 (usually
operating at 5 volts or less) from that elevated voltage, and to
enable G2S 842 to effectively control HS2S 838.
[0100] Low side capacitors C1P 1008, C2P 1018, C1S 1034 and C2S
1044 help reduce the resonance frequency of the converter 1000.
Maximum output power of the converter 1000 is a function of the
resonance frequency and transformer 806 self inductance. Capacitors
C1P 1008, C2P 1018, C1S 1034 can be physical capacitors, or can
correspond to parasitic capacitance, depending on the maximum
output power and the design of the transformer 806.
[0101] An NMOS transistor is turned on when its gate voltage is
sufficiently higher than its source voltage. The bootstrap
capacitors 1016, 1026, 1042, 1052 are used to elevate the gate
voltage provided by respective control signals 818, 820, 840, 842
to enable their respective high side PMOS transistors 814, 816,
836, 838 to turn on. As shown in FIG. 10, source voltages (V1P 822,
V2P 824) of input side 802 high side PMOS transistors 814, 816 can
reach approximately Vin 812; and source voltages (V1S 844 and V2S
846) of output side 804 high side PMOS transistors 836, 838 can
reach approximately Vout 822. Using bootstrap capacitors enables
high side PMOS transistor gate control voltages (outputs of level
shifting buffers 1016, 1026, 1042, 1052, which are dependent on
control voltages 818, 820, 840, 842) to reach higher than the input
voltage Vin 812 or output voltage Vout 834 (as appropriate).
[0102] The charging path for bootstrap capacitor C.sub.BS1P 1014
(path of current that charges the capacitor) runs from Vin 812,
through HS2P 816, to V2P 824, through VL1P 1002 and BS1P 1010, to
V.sub.BS1P 1012, to C.sub.BS1P 1014.
[0103] The charging path for bootstrap capacitor C.sub.BS2P 1024
runs from Vin 812, through HS1P 814, to V1P 822, through VL2P 1004
and BS2P 1020, to V.sub.BS2P 1022, to C.sub.BS2P 1024.
[0104] The charging path for bootstrap capacitor C.sub.BS1S 1040
runs from GND_S 856, through LS2S 850 to V2S 846, through VL1S 1028
and BS1S 1036, to V.sub.BS1S 1038, to C.sub.BS1S 1040.
[0105] The charging path for bootstrap capacitor C.sub.BS2S 1050
runs from GND_S 856, through LS1S 848 to V2S 844, through VL2S 1030
and BS2S 1046, to V.sub.BS1S 1048, to C.sub.BS1S 1050.
[0106] The bootstrap capacitors 1014, 1024, 1040, 1050 charge when
their respective adjacent node 822, 824, 844, 846 is low.
Accordingly, C.sub.BS1P 1014 charges when V1P 822 is low;
C.sub.BS2P 1024 charges when V2P 824 is low; C.sub.BS1S 1040
charges when V1S 844 is low; and C.sub.BS2S 1050 charges when V2S
846 is low.
[0107] The converter 1000 also preferably comprises an input
decoupling capacitor 858 connected between Vin 812 and GND_P 832,
and an output decoupling capacitor 860 connected between Vout 834
and GND_S 856 (as described hereinabove for decoupling capacitors,
also called ripple capacitors, of FIG. 2). Input and output
decoupling capacitors 858, 860 are not shown for simplicity.
[0108] The described embodiments provide one or more of at least
the following advantages. However, not all of these advantages
result from every one of the described embodiments, and this list
of advantages is not necessarily exhaustive. [0109] Enables higher
efficiency and decreased power use; [0110] zero voltage and zero
current switching; [0111] enables higher voltage input and output
with high efficiency; [0112] enables higher voltage output and/or
input without non-CMOS input and/or output capacitors; [0113]
enables integrated capacitances; [0114] reduces device area
requirements for input and/or output capacitors; [0115] enables use
of a low transformer turns ratio [0116] enables use of higher
switching frequencies; [0117] avoids driving losses and QG/RDS
tradeoff in self-driven transistors; [0118] enables use of
simplified rectification circuitry; [0119] enables use of
simplified driving control on the secondary side; and [0120]
reduces heat output.
[0121] In some embodiments, PMOS transistors cross-coupled between
different terminals of a primary or secondary side of a transformer
connect that side to a voltage input (on the primary side) or a
voltage output (on the secondary side). NMOS transistors couple
different terminals of that side of the transformer to the
respective (primary or secondary side) ground. This enables the
voltage input-connected transistors to be self-driven such that
gate charge is reused, turned into current across the respective
side of the transformer when the gate discharges. Further, NMOS
transistors are preferably zero current switched (ZCS) and zero
voltage switched (ZVS), avoiding significant switching losses.
Also, transistor configuration and switch timing avoids negative
input and output current. As a result, embodiments using high side
cross-coupled switches are highly efficient.
[0122] In some embodiments, high side NMOS transistors couple
different terminals of a primary or secondary side of a transformer
to a voltage input (on the primary side) or a voltage output (on
the secondary side). Low side NMOS transistors, cross-coupled
(through voltage limiters coupled to the low side NMOS transistors'
gates) to different terminals of the primary or secondary side of
the transformer connect that side to the respective (primary or
secondary side) ground. This enables the ground-connected
transistors to be self-driven such that gate charge is reused,
turned into current through the respective side of the transformer
when the gate discharges. Further, high side transistors are
preferably zero current switched (ZCS) and zero voltage switched
(ZVS), avoiding significant switching losses. Also, transistor
configuration and switch timing avoids negative input and output
current. As a result, embodiments using high side cross-coupled
switches are useful with high input and/or output voltages and are
highly efficient.
MODIFICATIONS AND VARIATIONS
[0123] This description has various references to a "zero" current,
a "zero" voltage, and a "zero" difference between voltages (e.g.,
"zero" voltage difference). As applied to example embodiments: (a)
"zero" means equal to 0 or else approximately 0; and (b)
"approximately 0" means near enough to 0 for the converter to
satisfy its efficiency specification, notwithstanding any remaining
losses (e.g., switching losses). In at least one example,
efficiency is a ratio between: (a) power received from the
converter; and (b) power applied to the converter. For an isolated
DC-DC converter of example embodiments, the converter's efficiency
specification is stated as a minimum threshold efficiency (such as
50%, which is a ratio of 1/2) for the converter to sustain during
its usual operation. The converter's efficiency specification is
one criteria to help determine whether the converter is suitable
for use within a particular operating environment. In some
examples, the transformer includes a magnetic coil, and the
converter's efficiency specification is 80% or higher.
[0124] In some embodiments as shown in FIGS. 2 and 6, biases and
sources of respective NMOS transistors are connected to each other,
such that NMOS transistors act as diodes, as shown in FIG. 5.
[0125] In some embodiments, an isolated DC-DC converter as shown in
FIG. 2, 4 or 5 can comprise NMOS instead of PMOS, and/or PMOS
instead of NMOS. In some embodiments, an isolated DC-DC converter
as shown in FIG. 2, 4 or 5 can comprise switches made using
technologies other than (or supplementary to) CMOS, e.g.,
LDMOS.
[0126] In some embodiments, a high input voltage and high output
voltage isolated DC-DC converter comprises an input side and an
output side, each side comprising a "stacked" configuration as
described with respect to the output side of FIG. 4. Accordingly,
the output side comprises a "stacked" output side as described with
respect to FIG. 4; and the input side comprises an input instance
and a ground instance of the input side of FIG. 2, the Vin (input
voltage) of the input instance comprising the Vin of the input side
(i.e., the voltage input for the converter), the ground of the
ground instance comprising the ground of the input side, and the
ground of the input instance connected to the Vin of the ground
instance. (Accordingly, the input side can comprise stacked
instances of the input side of, e.g., FIG. 2.)
[0127] In some embodiments as shown in FIGS. 4 and 5, isolation
capacitors isolate the ground instance from the transformer. In
some embodiments as shown in FIGS. 4 and 5, isolation capacitors
isolate both the ground instance and the output instance from the
transformer.
[0128] The output instance and ground instance are also called an
output module and ground module herein.
[0129] Modifications are possible in the described embodiments, and
other embodiments are possible, within the scope of the claims.
* * * * *